|
|
|---|
|
American Politics Journal Atrios Barry Crimmins Betty Bowers Buzzflash Consortium News Daily Howler Daily Kos Democatic Underground Disinfotainment Today Evil GOP Bastards Faux News Channel Greg Palast The Hollywood Liberal Internet Weekly Jesus General Joe Conason Josh Marshall Liberal Oasis Make Them Accountable Mark Morford Mike Malloy Political Humor - About.com Political Wire Randi Rhodes Rude Pundit Smirking Chimp Take Back the Media Whitehouse.org More Links |
Subject: the anti-Hillary poll on Kos
Greetings My Dear Bartcop:
You wrote::
> "Last time I checked it has 11,000 voters, which is kind of scientific.
If you ask 5 people who
> they're voting for - that's not science. The major polls only
ask about 1200 people, and they have
> a margin of error of about 3 %. But when you ask 11,000, that's
pretty damn scientific."
The power of good polling is not in the
number of people you ask. Rather, it all depends on
HOW you SELECT the people you survey. If you
were to ask 11,000 people at the Republican
National Convention who they think was the greatest
president ever - you will NOT get an accurate
reflection of who the entire US population thinks
was the greatest president. Similarly,the same
error would occur if you only asked folks at
the Democratic convention.
I see, but what if you asked 11,000 Republicans and Reagan
only got 1.5
percent of the vote.
Would you think the poll was rigged? I damn sure would.
The key is to RANDOMLY select a small sample of
people from the population that you wish to measure.
This should result in a REPRESENTATIVE SAMPLE
which statistically should give you an accurate
reflection of the larger population (with the
usual ± a few %) without having to ask ALL of them.
I get that, I really do, but I think you're helping me make my point.
The Kos poll is, presumably, mostly Democrats.
Why would the front runner have no representation there?
It's almost like the pro-Hillary Democrats (there must be some) couldn't
vote.
What would explain a lot is maybe Kos was kidding with this poll.
Back in 2001, we had a pretend comedy war with Bushwatch. (Scroll down a little)
Some fan of the page wrote a "script" that voted for bartcop.comhundreds
of times per hour.
I'm not suggesting Kos is crooked - I'm suggesting a hacked vote would
explain a lot.
It's the only thing that would explain the front runner getting fewer
votes than "I don't know."
As you can see from all the weekly polls, the
numbers of individuals actually polled are usually
between 1100 and 1300 people. When done correctly
this will result in an accurate representation
of what the entire population of the USA (all
300 million of us) thinks.
Picking the wrong group of people to ask resulted
in the famous "Dewey Defeats Truman"
prediction and headline in 1948. As I recall
the story, they took a telephone poll and Dewey
came out on top. However, in 1948 telephones
were much more of a sign of affluence than
they are today. Thus the results were skewed
toward rich folks who supported Dewey and
away from po folk who supported (and ultimately
voted for) Truman.
Lastly, any poll where those who are "polled"
are "self selected" rarely give meaningful results
that can be generalized to the larger population.
This is often why the candidates who are elected
in primary races (by a small self-selected group
of people who are very motivated by politics)
often encounter problems in the general election.
The views of the very motivated primary voters
are not proportionally reflected by the larger
population of voters (who are far less motivated
by politics) in general elections.
That's my two bits worth for tonight.
Keep up the great work,
Stan in Durant.
Stan, I get that - and what you say is valid.
What you're saying might explain things if Hillary came in second with
25 percent.
That would tell me the poll was "inaccurate."
But allow me to quote Brit Hum (R-Fascist dog)
Hume drearily opined:
"Readers of the Internet's largest and most
influential liberal Web blog overwhelmingly
want Al Gore in 2008. Sixty-eight percent
of respondents in a DailyKos.com survey
preferred Gore — who has said he's not considering
a run for the White House."
Quoting newshounds.com:
"Fox, which typically vilifies left-wing blogs,
was ecstatic to see Hillary show poorly
in multiple "left-wing" polls and used it
to their advantage to continue the campaign
of smearing her - since they are obviously
so frightened of the power she wields."
I wish Kos would help FOX News smear good Democrats less.